High rate of
PIK3CA
mutations but no
TP53
mutations in
low-grade adenosquamous carcinoma of the breast
Guillaume Bataillon,
1Laetitia Fuhrmann,
1Elodie Girard,
1Emanuelle Menet,
2Marick La
e,
1Mathieu Capovilla,
3Isabelle Treilleux,
4Laurent Arnould,
5Frederique Penault-Llorca,
6Roman Rouzier,
7Caterina Marchi
o,
1,8Ivan Bieche
1& Anne Vincent-Salomon
11Institut Curie, Paris Sciences Lettres Research University, P^ole de medecine diagnostique et theranostique, Paris, 2Institut Curie, Versailles Saint Quentin University, P^ole de medecine diagnostique et theranostique, Saint-Cloud, 3Department of Pathology, Centre Francßois Baclesse, Caen,4Department of Pathology, Centre Leon-Berard, Lyon, 5Department of Pathology, Centre Georges Francßois Leclerc, Dijon,6Department of Pathology, Centre Jean Perrin,
Clermont-Ferrand,7Institut Curie, Versailles Saint Quentin University, Surgery Department, Saint-Cloud, France, and
8Department of Medical Sciences, University of Turin, Turin, Italy
Date of submission 5 January 2018 Accepted for publication 8 March 2018 Published online Article Accepted 14 March 2018
Bataillon G, Fuhrmann L, Girard E, Menet E, Lae M, Capovilla M, Treilleux I, Arnould L, Penault-Llorca F, Rouzier R, Marchio C, Bieche I & Vincent-Salomon A
(2018) Histopathology 73, 273–283. https://doi.org/10.1111/his.13514
High rate of
PIK3CA
mutations but no
TP53
mutations in low-grade adenosquamous
carcinoma of the breast
Aims: Low-grade adenosquamous carcinoma of the breast (LGASC) is a rare variant of metaplastic carci-noma characterised by a favourable outcome and histo-logically composed of glandular and squamous elements in a spindle cell background typically associ-ated with a lymphocytic stromal reaction. Because of its rarity, the immunophenotypic and genetic profile of LGASC has not been sufficiently characterised. The aim of this study was to gain insights into the molecular and phenotypic characteristics of LGASC.
Methods and results: We reviewed the clinical and morphological features and detailed the immunohisto-chemical characteristics of a retrospective series of 13 LGASCs. Targeted sequencing of 50 genes was per-formed in 10 of 13 cases. Identified mutations were further assessed by Sanger sequencing in a validation series of 11 additional cases. All tumours showed a triple-negative immunophenotype, expressed ‘basal’
keratins, showed variable levels of epidermal growth factor receptor expression, and did not express androgen receptor. Sequencing analysis of the screening set of LGASCs revealed a high rate (seven of 10 cases) of PIK3CA mutations, whereas no TP53 mutations were found. All PIK3CA mutations were missense mutations located either in exon 20 (n = 6) or in exon 9 (n = 1). The global PIK3CA mutation rate, including the validation series, was 52% (11 of 21 cases). No disease recurrences were observed. [Correction added on 11 June 2018, after first online publication: The percentage of mutation rate was corrected to 52%]
Conclusions: Our results indicate that LGASC of the breast is a low-grade triple-negative breast cancer that harbours a basal-like phenotype with no andro-gen receptor expression, and shows a high rate of PIK3CA mutations but no TP53 mutations.
Keywords: androgen receptor, breast cancer, low-grade adenosquamous carcinoma, metaplastic carcinoma, PIK3CA mutations
Address for correspondence: Dr A Vincent-Salomon, Institut Curie, P^ole de medecine diagnostique et theranostique, 26 rue d’Ulm, 75248 Paris Cedex 05, France. e-mail: anne.salomon@curie.fr
Dr G Bataillon, Institut Curie, P^ole de medecine diagnostique et theranostique, 26 rue d’Ulm, 75248 Paris Cedex 05, France. e-mail: guillaume.bataillon@curie.fr
© 2018 The Authors. Histopathology published by John Wiley & Sons Ltd.
Introduction
Low-grade adenosquamous carcinomas of the breast (LGASCs) represent a rare histological type of invasive breast cancer that has been included in the latest edi-tion of the World Health Organisaedi-tion classificaedi-tion,1 as part of the spectrum of metaplastic carcinomas of the breast. They feature well-developed gland/tubule formation intimately admixed with solid nests of squamous cells in a spindle cell background. Rosen and Ernsberger2 emphasised that, despite the pres-ence of metaplastic elements, these tumours show a low-grade histological pattern.
Typically, the carcinomatous component is charac-terised by small glandular structures with rounded rather than angulated contours, and solid cords of epithelial cells that may contain squamous cells, squamous pearls, or squamous cyst formation. The epithelial structures of adenosquamous carcinomas are often dispersed in a highly cellular spindle cell background. These spindle cells appear to merge with the epithelial cells and, in some cases, spindle cell metaplasia may be identified. Some authorities have suggested that the spindle cell background may be considered as part of the neoplastic component.3 However, it may sometimes be difficult to distinguish between spindle cell metaplasia and ordinary stroma,4,5and this fusiform background may be chal-lenging to interpret in core needle biopsy specimens. In agreement with their low-grade morphological fea-tures, the majority of LGASCs show an excellent prognosis with a low incidence of lymph node metastasis.6 A proportion of cases, however, can behave in a locally aggressive manner.6
Although several case reports and relatively small cohorts have been reported over the years, LGASC remains poorly defined in terms of phenotype; in addition, its genetic profile is yet to be established.3,6–
16The immunohistochemical studies published to date
have described LGASC as part of the triple-negative breast cancer (TNBC) spectrum.1,6,17,18
A genome-wide characterisation of copy number alterations has been performed on two cases, one of which showed a complex profile of copy number changes encompassing focal amplifications.3 A thor-ough genomic characterisation of metaplastic carci-nomas has been recently provided,19 but, regrettably, LGASCs were not present in the cohort.
To gain greater insights into the specific biological characteristics of LGASC, in this study we report a detailed morphological and immunophenotypic analy-sis of 13 cases and the first characterisation of the
genomic landscape of LGASC by use of a targeted panel of genuine cancer genes.
Materials and methods
C A S E SWe retrospectively retrieved 13 LGASCs diagnosed at Institut Curie (Paris and Saint Cloud) and at Centre Francßois Baclesse (Caen) between 1999 and 2013. These carcinomas were originally registered as either ‘adenosquamous’ or ‘metaplastic’ carcinomas. Sam-ples were anonymised before analysis. Ethical approval from the Institutional Review Board (Institut Curie breast cancer study group: IRB 03/14) was obtained for the use of all specimens. Clinical data on treatment and follow-up were obtained from the Insti-tut Curie electronic medical records.
A second independent cohort of 11 LGASCs that served as a validation series was obtained from the Centre Georges Francßois Leclerc in Dijon, Centre Jean Perrin in Clermont-Ferrand, and Centre Leon-Berard in Lyon.
M O R P H O L O G I C A L A N A L Y S I S
Histopathological review was performed by two expe-rienced breast pathologists (A.V.S. and G.B.) for the purpose of subtyping and grading (using the Elston– Ellis grading system).20 Specific histological features commonly seen in LGASC, such as degree of glandu-lar, squamous and spindle cell differentiation, were assessed, and tumours were classified according to the predominant component (glandular, squamous, and spindle cell). In addition, stromal characteristics, including the presence/degree of lymphoid aggre-gates, were recorded. The respective proportions of glandular and squamous tumour cells and the spindle cell component, lymphoid stromal cells and associated normal glands within the tumour were determined. The presence of lymphovascular invasion and lymph node metastasis was also recorded.
I M M U N O H I S T O C H E M I S T R Y
Three-micrometre-thick formalin-fixed paraffin-embedded (FFPE) sections were mounted on silane-coated slides. Immunohistochemical staining was performed with 16 different antibodies directed against oestrogen receptor (ER), progesterone receptor (PR), androgen receptor (AR), human epidermal growth factor receptor 2 (HER2), Ki67, epidermal
growth factor receptor (EGFR), p63, cytokeratin (CK) 5/6, CK7, CK14, CK8/18, a-smooth muscle actin, E-cadherin, CD117, PTEN, and EZH2. Details of anti-body sources, dilutions and antigen retrieval methods are summarised in Table S1. Positive and negative controls (omission of the primary antibody and IgG-matched serum) were utilised.
All markers were scored (percentage of positive tumour cells) separately in the glandular, squamous and spindle cell components of the tumours. ER, PR and HER2 were evaluated according to the current American Society of Clinical Oncology/College of American Pathologists guidelines.21,22 The positive cut-off for expression of all other markers was 10%. Markers were characterised according to the highest percentage of positive cells within each component.
M A C R O D I S S E C T I O N A N D D N A E X T R A C T I O N
All tumours were macroscopically dissected to enrich for tumour cell content, by taking out normal breast tis-sue and any areas of fibrocystic changes and intraduc-tal papillomas. Macrodissection was performed with a sterile scalpel from three consecutive 6-lm-thick sec-tions of FFPE samples after deparaffinisation, guided by the microscopic limits of the tumour on the haema-toxylin and eosin-stained section. DNA was extracted with the NucleoSpin 8 tissue Kit (Macherey Nagel, D€uren, Germany), according to the manufacturer’s rec-ommendations. The DNA level was measured with the Qubit 2.0 fluorometer assay according to the manufac-turer’s instructions (Life Technologies, Thermo Fisher Scientific, Waltham, MA, USA). A KRAS polymerase chain reaction (PCR) was performed with LightMix Kit Kras to assess DNA quality. Following DNA quantifica-tion and quality assessment, three of 13 cases were withdrawn because of lack of dissected material (case 6) or excessively low DNA levels (cases 2 and 4).
T A R G E T E D M A S S I V E P A R A L L E L S E Q U E N C I N G ( M P S )
Given that only FFPE material was available, we focused on mutations affecting known oncogenes and tumour suppressor genes in human cancers. For this purpose, targeted sequencing was performed with the Ion Ampliseq Cancer Hotspot panel v2 in conjunction with the Ampliseq library kit v2.0 and the Ion Tor-rent Personal Genome Machine (PGM) Analyzer (Life Technologies, Carlsbad, CA, USA). According to the manufacturer’s instructions, 10 ng of extracted DNA was used to generate 207 amplicons surveying 2800 COSMIC mutations in 50 well-established oncogenes
and tumour suppressor genes that are frequently mutated in different malignancies, including breast cancer (summarised in Table S2).
Extremities of amplicons were partially digested, and Ion adapters (Ion Xpress PCR barcode adapters; Life Technologies), including one with a molecular barcode, were ligated at both ends. After amplifica-tion of the construcamplifica-tion, the quality of the libraries was checked on a BioAnalyzer (Agilent, Santa Clara, CA, USA) and quantified with Qubit Technology (Life Technologies). Template preparation was performed with the Ion OneTouch System 2 with the Ion PGM Template OT2 200 Kit (Life Technologies). Templates were sequenced on the Ion Torrent PGM with an Ion PGM Sequencing 200 Kit v2 and an Ion 318v2 Chip. The overall quality of each run was evaluated on the basis of the report generated by the Torrent Ser-ver. At least 100 000 reads per sample were required. A sample was considered to be valid only if 95% of targeted positions were covered at >1009, according to the analysis performed by the variant caller plugin ofTORRENT SUITE (Life Technologies).
V A R I A N T C A L L I N G
Data from the PGM runs were initially processed with the Ion Torrent platform-specific pipeline software
TORRENT SUITEto generate sequence reads, trim adapter
sequences, and filter and remove poor signal-profile reads. Initial variant calling from the Ion AmpliSeq sequencing data was generated by TORRENT SUITE v3.0 with a plug-in ‘variant caller’ program. For mutation analyses, sequencing reads were aligned on the human reference genome (hg19) by the use of TMAP (Life Technologies). Variants were detected with
VARIANT CALLER (Life Technologies), and annotated by
use of the ANNOVAR pipeline.23 Variants were fil-tered according to their frequency [>4% for single-nucleotide variants (SNVs) and >10% for indels], strand ratio (>0.2), and read coverage (>309 for SNVs and>1009 for indels). The Bioinformatics plat-form integrated the different molecular results in a name-blinded technical report, which was then discussed and interpreted by a pathologist and a molecular biologist together (G.B. and I.B.). Align-ments were verified visually with the Integrative Genomics Viewer v.2.2 (Broad Institute).
S T A T I S T I C A L A N A L Y S I S
COSMIC (version 64) and the My Cancer Genome database (http://www.mycancergenome.org/) were used to assess recurring mutations in breast cancer.
Statistical correlations were performed with GRAPHPAD
PRISM version 6 (GraphPad Software, La Jolla, CA,
USA).
V A L I D A T I O N O F M U T A T I O N S
Mutations identified in the 10 cases subjected to tar-geted MPS were validated by Sanger sequencing (LightCycler 480 Real-Time PCR system, kit PIK3CA, Roche Molecular Systems, Indianapolis, IN, USA). The frequency of PIK3CA mutations was further explored by Sanger sequencing, including in the vali-dation series of 11 LGASCs (Table S3).
Results
C L I N I C A L F E A T U R E S
The clinicopathological features of the cohort are sum-marised in Table 1. All patients were female, and they were predominantly of postmenopausal age. Tumour size ranged from 3.5 mm to 70 mm (mean, 22 mm; median, 15 mm). One case (case 8) had axillary lymph node metastases showing a ductal morphology remi-niscent of a previously diagnosed contralateral breast carcinoma (Table 1), and showed ER and PR expres-sion. Clinical follow-up (range, 3–17 years; mean, 7.5 years) did not reveal any recurrences.
H I S T O P A T H O L O G Y
As summarised in Table 2, all cases had varying degrees of glandular, squamous and spindle cell com-ponents (Figure 1).
The epithelial component was organised in small tubules, cell clusters and cords with focal squamous differentiation in a spindle cell background. Neoplastic glands were elongated, ovoid, or compressed, and some infiltrated in and around residual normal glands. Spindle cells and stromal cells (fibroblasts and lymphoid cells) intimately surrounded the epithelial cell nests. Lymphocytic aggregates were present in all tumours. Neither necrosis nor vascular invasion was observed.
Case 3 showed a focal high-grade component, in line with previous descriptions of progression from low-grade adenosquamous carcinoma to high-grade metaplastic carcinoma.2,24
I M M U N O P H E N O T Y P I C A N A L Y S I S
Immunohistochemistry results are summarised in Figures 2 and 3 and detailed in Table S4. All
tumours were of the triple-negative phenotype (ER-negative, PR-negative, and HER2-negative) and lacked AR expression. The proliferation indices assessed according to Ki67 ranged between 2% and 20% (mean, 11%; median, 10%). The immunophe-notype was found to be relatively similar across the various cases, depending on their predominant com-ponent: glandular, squamous, or spindle cell (Figure 2).
The glandular component was characterised by dif-fuse (>50%) expression of CK7 and CK14, E-cadherin, and EZH2. The squamous component expressed CK14, E-cadherin, CK7, CK5/6, EZH2, p63, and PTEN, whereas the spindle cell component expressed SMA, PTEN and, to a lesser extent, EZH2. Some of the spindle cells showed expression of CK5/6, CK14, CK7, and p63.
M U T A T I O N F R E Q U E N C Y I N K N O W N C A N C E R G E N E S
Mutations affecting the 50 genes included in the targeted AmpliSeq Panel V2 are summarised in Table 2 and detailed in Table S5. Ten breast LGASC samples were analysed by MPS. A mean 1009 coverage of 90% with a mean read length of 226 bp was obtained in the 10 FFPE samples. The majority of the samples (seven of 10 cases, 70%) harboured missense PIK3CA mutations affecting exon 20 (n = 6) or exon 9 (n = 1), and corresponding to the kinase (H1047L and H1047R) and helical (E542K) domains, respec-tively. Two of the samples showing a PIK3CA mutation also harboured mutations in GNAS and KIT (cases 1 and 3, respectively). Co-occurrence of CDKN2A and PTEN mutations was detected in one other case (case 7). TP53 was sequenced in all exons except exon 1, and no TP53 mutations were identified.
Six of the seven PIK3CA mutations identified by Ion Torrent sequencing analysis were subsequently validated by Sanger sequencing in the index cases. It should be noted that the PIK3CA mutation not detected by Sanger sequencing showed an allele fre-quency of 6.7% in the Ion Torrent sequencing assay. To confirm the high rate of PIK3CA muta-tions in this entity, we analysed an independent series of 11 LGASCs and found four tumours (36%) harbouring PIK3CA mutations in exon 20 (n= 2) or exon 9 (n = 2) (Table S5B). The global PIK3CA mutation rate, including screening and validation series, accounted for 52% (11 of 21 cases) (Table S5C).
Table 1. Clinical characteristics of the cohort of 13 low-grade adenosquamous carcinomas of the breast (LGASCs) Case Ag e (y ears) Prev ious brea st lesion Famil y history of cance r In itial clinical prese ntation Size (m m) Lymp h node met astasis M astecto my Chemo therapy Rad iothera py Dise ase-free surviva l (years) 1 5 0 Benign None Palp able mass 25 0/1 4 T o tal No No 17 2 5 4 None Moth er (breas t) Palp able mass 20 NA Parti al No Yes 13 3 5 4 None Fath er (oesop hagu s) Sc reening mammo graphy 10 0/1 0 Parti al Yes Yes 11 4 8 5 None None Palp able mass 12 NA Parti al No No 11 5 5 4 None Fath er (colon ) Palp able mass 15 0/4 Parti al No Yes 9 6 5 6 Invasive lobul ar carc inoma None Sc reening mammo graphy 3. 5 0/1 Parti al No Yes 8 7 6 6 None None Palp able mass 20 0/1 9 Parti al No Yes 6 8 8 1 Invasive duc tal carc inoma Moth er (breas t) Palp able mass 35 10/ 15* Parti al Yes Yes 5 9 4 6 None None Palp able mass 34 NA Parti al No No 5 10 66 None None Sc reening mammo graphy 10 0/1 2 Parti al Yes Yes 4 11 62 None None Sc reening m ammo graphy 14 0/1 9 Parti al Yes Yes 3 12 58 Benign Moth er (breas t) Palp able mass 70 0/1 0 Parti al Yes Yes 3 13 28 None Moth er (breas t) Palp able mass 11 0/2 Parti al No Yes 3 NA, not avail able. *Lym ph node metastas is of contra lateral oestrog en recep tor-posit ive invas ive duct al carc inoma (i. e. invas ive car cinom a o f n o spec ial ty pe).
Table 2. Pathological features of 13 low-grade adenosquamous carcinomas of the breast (LGASCs) Cases Histol ogical grade* Tum our comp onen ts Stroma Normal glands (%) † Other asso ciated lesion s Muta ted genes ‡ Gland formati on Nuclear atypia Mitoses Fi nal grad e Gla ndular cel ls (%) † Squamo us cells (%) † Spindle cells (%) † Fibroblasts (%) † Lymp hoid cells (%) † Lympho id infiltrate 1 3 1 2 II 5 3 0 1 0 5 25 Intrat umour 25 Intradu ctal papil loma PIK3 CA , GNAS 2 3 1 1 I 0 30 15 10 15 Intrat umour 30 – NA 3 3 2 1 II 40 10 15 5 2 0 Intrat umour 10 Fibro cystic disease, papil loma PIK3 CA , KIT 4 2 2 1 I 1 5 1 5 4 0 5 5 Intrat umour 20 – NA 5 2 2 1 I 1 5 3 0 2 5 1 5 1 0 Intrat umour 5 Fibro adenoma PIK3 CA 6 3 2 1 II 10 25 40 10 5 Periphe ral 10 – NA 7 3 2 1 II 10 10 25 5 1 5 Intrat umour 40 Fibro cystic disease, papil loma CDK N2A , PTEN 8 3 3 1 II 25 15 35 10 5 Intrat umour 10 Fibro cystic disease, papil loma PIK3 CA 9 2 2 1 I 5 5 4 0 1 0 2 0 Intrat umour 20 – PIK3 CA 10 3 2 1 II 1 5 3 0 3 0 1 0 1 0 Intrat umour 5 Aden osis PIK3 CA 11 2 2 1 I 10 10 30 0 1 0 Periphe ral 40 Fibro adenoma Wil d ty p e 12 2 1 1 I 10 10 15 10 15 Intrat umour 40 Intradu ctal papilloma Wil d ty p e 13 2 1 1 I 25 10 15 30 10 Intrat umour 10 Fibro cystic breast disea se PIK3 CA NA, not avail able. *Acc ording to Elston and Ellis. 20 †The rela tive perce ntag es of distinct cel l comp onen ts of the lesion s sum up to 100% . ‡The mutati onal st atus reported here incl udes details on the hotspot mutatio ns investig ated in 50 cance r genes.
Discussion
In this study, we show that LGASC represents an indolent form of TNBC with basal-like features char-acterised by high rates of PIK3CA mutations and lack of TP53 mutations. These data on LGASC provide insights into the genomic landscape of metaplastic breast cancer, and may help in the understanding of the possible evolutionary pathways of low-grade forms of TNBC, by representing a starting point for future studies of putative precursors of LGASC and other metaplastic carcinomas. These findings may also be of help in the diagnostic setting, because, although LGASC can be relatively easily recognised on morphological grounds in surgical specimens, it may be a challenging entity to be diagnosed on core biopsy samples with respect to other fusiform prolifer-ations, especially those of low grade.
LGASC is an unusual form of metaplastic carci-noma featuring a variable admixture of well-formed glands and solid cords of cells with squamous features arranged in a haphazard, infiltrative pattern and typi-cally showing relatively little cytological atypia and low levels of proliferation, to the extent that differen-tial diagnosis with radial scar (RS)/complex sclerosing lesions (CSLs) can be challenging. The extension of
the lesion in a spindle cell background that can stand for the presence of desmoplastic stroma and the pres-ence of a triple-negative phenotype may represent helpful features for the differential diagnosis.
So far, relatively small cohorts characterising this entity have been reported, and almost exclusively at the immunophenotypic level. In agreement with pre-vious studies,6,18 all of the cases pertaining to our series were of the triple-negative phenotype. In addi-tion, AR expression was not observed. This is in con-trast to the report of the large majority of metaplastic carcinomas being AR-positive, except for squamous cell carcinomas.25We also confirmed in the epithelial neoplastic component the consistent expression of basal CKs and of markers usually expressed in normal myoepithelial cells. In a subgroup of cases, a minor proportion of the spindle cells also expressed p63, CKs, and EZH2, which is a protein that is overex-pressed in a wide range of solid tumour types, includ-ing breast cancer.26 Taken together, these preliminary observations may indirectly suggest that the spindle cell background observed in LGASC could be an integral part of the neoplastic process. This hypothesis should be further explored in studies aimed at analysing the molecular features of the dis-tinct cell components of LGASCs. Interestingly Geyer
A B
C D
E F
Figure 1. Histological features of low-grade adenosquamous carcinomas of the breast (Hematoxylin-Eosin-Saffron). This entity is composed of infiltrating glandular structures (A) with variable degrees of squamous differentiation (B,C) in a spindle cell background (D) entrapping residual normal glands (C–E). Lymphocytic aggregates were present in all tumours (D–F).
et al.3 have reported on the molecular karyotype obtained in two cases of LGASC by microarray com-parative genomic hybridisation, and have shown a rather simple genome in one case and a higher degree of complexity of copy number changes in the other case, which harboured a focal EGFR amplifica-tion. By means of chromogenic in-situ hybridisation analysis, the authors were able to demonstrate that the EGFR amplification found in the epithelial cells
was also present in a minority of the spindle cells of the stromal component, in particular in those near the epithelial clusters.
With the exception of the study above, the molecu-lar underpinning of LGASC is molecu-largely unknown. The advent of MPS has contributed to a thorough charac-terisation of the genetic landscape of breast cancer.27 Recently, it has been demonstrated that, although metaplastic carcinomas and common forms of TNBC
80% 70% 90% 70% 60% 0% 0% 0% 40% 50% 50% 60% CK7 n = 10/11 CK14 n = 9/9 E-Cad n = 8/8 EZH2 n = 7/9 C-kit n = 6/10 SMA n = 9/11 P63 n = 10/10 CK8/18 n = 5/10 PTEN n = 6/10 CK5/6 n = 8/11 EGFR n = 4/11 CK7 n = 8/12 CK14 n = 4/10 E-Cad n = 1/10 EZH2 n = 8/10 C-kit n = 0/11 SMA n = 13/13 P63 n = 11/13 CK8/18 n = 1/11 PTEN n = 10/11 CK5/6 n = 6/12 EGFR n = 4/12 CK7 n = 8/10 CK14 n = 9/9 E-Cad n = 8/9 EZH2 n = 9/10 C-kit n = 2/10 SMA n = 9/11 P63 n = 11/11 CK8/18 n = 1/10 PTEN n = 10/10 CK5/6 n = 8/8 EGFR n = 3/10 80% 65% 60% 45% Glandular Squamous Spindle 10% 5% 3% 30% 0% 0% 8% 0% 90% 10% 70% % = median values 0–40% 40–80% ≥ 80% 40% 35% 18% 15% 10% 0%
Figure 2. Immunophenotype of glandular, squamous and spindle cell components. Median values of percentages of positive cells (grey bar plots) within the glandular, squamous and spindle cell components are shown. Numbers of positive cases relative to the total number of cases available for analysis per marker (the threshold for positivity was set at 10%) are indicated at the bottom of the columns.
harbour similar mutational frequencies for TP53, they are genetically distinct entities.19 Metaplastic carcinomas show more frequent mutations in PIK3CA, PIK3R1, PTEN (57% of cases), and Wnt pathway genes (51% of cases). PIK3CA mutations are significantly more frequently observed in spindle cell metaplastic carcinomas than in other metaplastic car-cinomas.19 In this context, our results provide another piece of evidence for heterogeneity within the spectrum of metaplastic carcinomas. We observed that LGASCs seem to lack TP53 mutations and to harbour a relatively high frequency of PIK3CA muta-tions (52%), in stark contrast to high-grade spindle cell metaplastic carcinomas, which harbour both TP53 and PIK3CA mutations.
Recently Geyer et al.28,29have reported on the exis-tence of a ‘low-grade triple-negative breast neoplasia family’ featuring microglandular adenosis/atypical microglandular adenosis as non-obligate precursors of TNBC, and acinic cell carcinomas as low-grade forms
of TNBC with the potential to transform into conven-tional high-grade TNBCs. Both microglandular adeno-sis/atypical microglandular adenosis associated with a carcinoma and acinic cell carcinomas have been shown to harbour TP53 mutations, and the acquisition of a TP53 mutation seems to represent the key event in the possible progression of microglandular adenosis/ atypical microglandular adenosis to a carcinoma.29In this respect, the lack of TP53 mutations and the pres-ence of PIK3CA mutations in LGASC may suggest that these lesions follow a molecular evolutionary pathway that is distinct from that of conventional TNBC. Inter-estingly, lesions such as papillomas,30 RS/CSLs31,32 and infiltrating epitheliosis33have also been shown to harbour PIK3CA mutations, and the causative role of PIK3CA mutations in these entities remains unclear. As a further level of complexity, an association between RS/CSLs and LGASCs has also been docu-mented.24,34 In this scenario, one could hypothesise that LGASC of the breast may represent an indolent
A B
C D
E F
Figure 3. Immunophenotypic profile of low-grade adenosquamous carcinomas of the breast (LGASCs). A, LGASC with glandular, squamous and spindle components (Hematoxylin-Eosin-Saffron). B, LGASC is a triple-negative carcinoma showing cytokeratin (CK) 5/6 expression. C, The spindle cells are positive fora-smooth muscle actin. D–F, The glandular and squamous components express CK14 (D), p63 (E), and EZH2 (F).
carcinomatous lesion whose non-obligate precursor may be identified in RS/CSLs, at least in a subgroup of cases. To unravel the causative role within this associ-ation, comparative molecular analyses of synchronous RS/CSLs and LGASCs are warranted. At present, a sin-gle report has documented a case with synchronous CSL featuring infiltrating epitheliosis and an LGASC that shared the same PIK3CA and SF3B1 mutations, thus providing evidence that the LGASC most likely originated from the CSL.33
Another low-grade variant of metaplastic carci-noma, i.e. fibromatosis-like metaplastic carcicarci-noma, may arise in direct association with sclerosing papil-lary lesions.34Molecular studies dissecting the genetic landscape of this lesion would be of help in ascertain-ing whether or not these lesions share a genetic back-ground with LGASC.
Our study has some limitations. The first is the lim-ited number of cases analysed by MPS; nevertheless, it has to be acknowledged that LGASC represents a rare entity, and the major finding of the sequencing related to PIK3CA mutations was validated in an independent cohort. Second, the targeted panel used was composed of a limited number of genes, and covered either hot-spots or a limited number of exons. It is rather remark-able, however, that none of the cases showed TP53 mutations, as this gene was fully covered in the sequencing assay except for exon 1. Notwithstanding this, we cannot exclude the possibility that more com-prehensive analyses in larger series may add to the real frequency of PIK3CA mutations, or unveil the presence of TP53 mutations in a subgroup of cases and provide evidence of additional recurrent mutations characteris-ing this entity. Finally, four of the 13 patients were fol-lowed up for <5 years, a time interval that may be suboptimal for investigating the behaviour of a low-grade form of invasive carcinoma of the breast. Never-theless, no disease recurrence events were observed in patients with longer follow-up.
However, even if we take these limitations into account, our study not only confirms the morphologi-cal and immunohistochemimorphologi-cal data already reported in the literature concerning this entity, but also provides evidence suggesting that LGASCs constitute a distinct genetic entity among metaplastic carcinomas, charac-terised by high rates of PIK3CA mutations and a lack of TP53 mutations, a triple-negative phenotype, and no AR expression. Although further studies are needed to confirm and expand our findings, this observation seems to corroborate the concept that LGASC should be distinguished from high-grade forms of metaplastic carcinoma and other triple-negative/basal-like carcino-mas that are associated with an aggressive behaviour.
At present, the clinical management of LGASC requires complete local and regional treatment, including sur-gery and radiotherapy. In our experience, axillary lymph node dissection and adjuvant therapies do not appear to be warranted, because of the favourable prognosis of these cancers. Additional analyses are required to assess the inflammatory stromal reaction in terms of antitumour immunity. Finally, further studies analysing the distinct cell components of LGASC are warranted to gain insights into the neoplastic nature of the spindle cells constituting the background of these tumours.
Acknowledgements
C. Marchio is supported in part by a grant from the Mayent-Rothschild foundation. The authors would like to thank Anthony Saul for help with the use of English.
Conflicts of interest
The authors have no conflicts of interest to declare.
Author contributions
G. Bataillon characterised the cohort at the morpholog-ical and immunophenotypic levels, performed experi-ments, analysed data, and wrote the manuscript. L. Fuhrmann curated the immunophenotypic characteri-sation of the cohort and participated in data analysis. E. Girard analysed molecular data. E. Menet and M. Lae participated in cohort review and in the interpreta-tion of immunohistochemical results. M. Capovilla, I. Treilleux, L. Arnould and F. Penault-Llorca provided samples and reviewed cases. R. Rouzier performed the clinical data analysis. C. Marchio participated in data analysis and wrote the manuscript. I. Bieche performed the molecular analysis, analysed data, and critically reviewed the manuscript. A. Vincent-Salomon con-ceived and supervised the study, provided samples, interpreted results, and wrote the manuscript.
References
1. Lakhani SR, Ellis IO, Schnitt SJ, Hoon Tan P, Van de Vijver M eds. World Health Organization classification of tumors of the breast. Lyon: IARC Press, 2012.
2. Rosen PP, Ernsberger D. Low-grade adenosquamous carci-noma. A variant of metaplastic mammary carcicarci-noma. Am. J. Surg. Pathol. 1987; 11; 351–358.
3. Geyer FC, Lambros MB, Natrajan R et al. Genomic and immunohistochemical analysis of adenosquamous carcinoma of the breast. Mod. Pathol. 2010; 23; 951–960.
4. Oh EY, Collins LC. Keratin expression patterns in stromal cells of benign sclerosing lesions of the breast: a potential diagnostic pitfall. Arch. Pathol. Lab. Med. 2015; 139; 1143–1148. 5. Rakha EA, Badve S, Eusebi V et al. Breast lesions of uncertain
malignant nature and limited metastatic potential: proposals to improve their recognition and clinical management. Histopathology 2016; 68; 45–56.
6. Soo K, Tan PH. Low-grade adenosquamous carcinoma of the breast. J. Clin. Pathol. 2013; 66; 506–511.
7. Boecker W, Stenman G, Loening T et al. Differentiation and his-togenesis of syringomatous tumour of the nipple and low-grade adenosquamous carcinoma: evidence for a common origin. Histopathology 2014; 65; 9–23.
8. Woodard BH, Brinkhous AD, McCarty KS Sr, McCarty KS Jr. Adenosquamous differentiation in mammary carcinoma: an ultrastructural and steroid receptor study. Arch. Pathol. Lab. Med. 1980; 104; 130–133.
9. Van Hoeven KH, Drudis T, Cranor ML, Erlandson RA, Rosen PP. Low-grade adenosquamous carcinoma of the breast. A clinocopathologic study of 32 cases with ultrastructural analy-sis. Am. J. Surg. Pathol. 1993; 17; 248–258.
10. Drudis T, Arroyo C, Van Hoeven K, Cordon-Cardo C, Rosen PP. The pathology of low-grade adenosquamous carcinoma of the breast. An immunohistochemical study. Pathol. Annu. 1994; 2; 181–197. 11. Agrawal A, Saha S, Ellis IO, Bello AM. Adenosquamous
carci-noma of breast in a 19 years old woman: a case report. World J. Surg. Oncol. 2010; 8; 44.
12. Noel JC, Buxant F, Engohan-Aloghe C. Low-grade adenosqua-mous carcinoma of the breast—a case report with a BRCA1 germline mutation. Pathol. Res. Pract. 2010; 206; 511–513. 13. Scali EP, Ali RH, Hayes M, Tyldesley S, Hassell P. Low-grade
adenosquamous carcinoma of the breast: imaging and histopathologic characteristics of this rare disease. Can. Assoc. Radiol. J. 2013; 64; 339–344.
14. Cha YJ, Kim GJ, Park BW, Koo JS. Low-grade adenosquamous carcinoma of the breast with diverse expression patterns of myoepithelial cell markers on immunohistochemistry: a case study. Korean J. Pathol. 2014; 48; 229–233.
15. Wilsher MJ. Adenosquamous proliferation of the breast and low grade adenosquamous carcinoma: a common precursor of an uncommon cancer? Pathology 2014; 46; 402–410. 16. Senger JL, Meiers P, Kanthan R. Bilateral synchronous
low-grade adenosquamous carcinoma of the breast: a case report with review of the current literature. Int. J. Surg. Case Rep. 2015; 14; 53–57.
17. Kawaguchi K, Shin SJ. Immunohistochemical staining charac-teristics of low-grade adenosquamous carcinoma of the breast. Am. J. Surg. Pathol. 2012; 36; 1009–1020.
18. Rakha EA, Coimbra ND, Hodi Z, Juneinah E, Ellis IO, Lee AH. Immunoprofile of metaplastic carcinomas of the breast. Histopathology 2017; 70; 975–985.
19. Ng CKY, Piscuoglio S, Geyer FC et al. The landscape of somatic genetic alterations in metaplastic breast carcinomas. Clin. Can-cer Res. 2017; 23; 3859–3870.
20. Elston CW, Ellis IO. Pathological prognostic factors in breast cancer. I. The value of histological grade in breast cancer: experience from a large study with long-term follow-up. Histopathology 1991; 19; 403–410.
21. Hammond ME, Hayes DF, Dowsett M et al. American Society of Clinical Oncology/College of American Pathologists guideline recommendations for immunohistochemical testing of estrogen and progesterone receptors in breast cancer. J. Clin. Oncol. 2010; 28; 2784–2795.
22. Wolff AC, Hammond ME, Hicks DG et al. Recommendations for human epidermal growth factor receptor 2 testing in breast cancer: American Society of Clinical Oncology/College of American Pathologists clinical practice guideline update. J. Clin. Oncol. 2013; 31; 3997–4013.
23. Wang K, Li M, Hakonarson H. Annovar: functional annotation of genetic variants from high-throughput sequencing data. Nucleic Acids Res. 2010; 38; e164.
24. Denley H, Pinder SE, Tan PH et al. Metaplastic carcinoma of the breast arising within complex sclerosing lesion: a report of five cases. Histopathology 2000; 36; 203–209.
25. Collins LC, Cole KS, Marotti JD, Hu R, Schnitt SJ, Tamimi RM. Androgen receptor expression in breast cancer in relation to molecular phenotype: results from the nurses’ health study. Mod. Pathol. 2011; 24; 924–931.
26. Wassef M, Michaud A, Margueron R. Association between EZH2 expression, silencing of tumor suppressors and disease outcome in solid tumors. Cell Cycle 2016; 15; 2256–2262. 27. Ng CK, Schultheis AM, Bidard FC, Weigelt B, Reis-Filho JS.
Breast cancer genomics from microarrays to massively parallel sequencing: paradigms and new insights. J. Natl Cancer Inst. 2015; 107; djv015.
28. Geyer FC, Pareja F, Weigelt B et al. The spectrum of triple-negative breast disease: high- and low-grade lesions. Am. J. Pathol. 2017; 187; 2139–2151.
29. Geyer FC, Berman SH, Marchio C, et al. Genetic analysis of microglandular adenosis and acinic cell carcinomas of the breast provides evidence for the existence of a low-grade triple-negative breast neoplasia family. Mod. Pathol. 2017; 30; 69– 84.
30. Troxell ML, Levine J, Beadling C et al. High prevalence of PIK3CA/AKT pathway mutations in papillary neoplasms of the breast. Mod. Pathol. 2010; 23; 27–37.
31. Wolters KL, Ang D, Warrick A, Beadling C, Corless CL, Troxell ML. Frequent PIK3CA mutations in radial scars. Diagn. Mol. Pathol. 2013; 22; 210–214.
32. Wilsher MJ, Owens TW, Allcock RJ. Next generation sequenc-ing of the nidus of early (adenosquamous proliferation rich) radial sclerosing lesions of the breast reveals evidence for a neoplastic precursor lesion. J. Pathol. Clin. Res. 2017; 3; 115– 122.
33. Eberle CA, Piscuoglio S, Rakha EA et al. Infiltrating epitheliosis of the breast: characterization of histological features, immunophenotype and genomic profile. Histopathology 2016; 68; 1030–1039.
34. Gobbi H, Simpson JF, Jensen RA, Olson SJ, Page DL. Metaplas-tic spindle cell breast tumors arising within papillomas, com-plex sclerosing lesions, and nipple adenomas. Mod. Pathol. 2003; 16; 893–901.
Supporting Information
Additional Supporting Information may be found in the online version of this article:
Table S1. List of antibodies used.
Table S2. Ion Ampliseq cancer Panel target genes. Table S3. Primers used for Sanger sequencing. Table S4. Immunophenotypic analysis of the dis-tinct components.
Table S5. Summary of PIK3CA mutations in LGASC (MPS and Sanger sequencing).
